Occupied with a clash of symbols

No writing system has been written about so pejoratively as Japanese,'' wrote a linguist recently. But, precisely the scandal of the Japanese writing system is also its primary fascination for all those interested in the ingenuity of writing. Japanese writing uses Chinese characters (kanji) that are essentially logographic symbols read in several Japanese or Chinese-derived pronunciations depending on how they function as elements in long- or short-word clusters. In addition, Japanese is also written with two complimentary syllabaries, katakana and hiragana. Both represent the sounds of Japanese with single signs for syllables, but katakana serves exclusively to transliterate foreign loan words such as pan ("bread"). Unlike kanji, there exists hardly any ambiguity in the pronunciation of words written with the kana syllabaries. The various readings of kanji can be rewritten in hiragana, and this is an increasingly common resort for the less used of the kanji. Finally, recent immigres to the Japanese script are the letters of the Roman alphabet, romaji, which can be used to transliterate the standard pronunciation of any Japanese word.

Literacy and Script Reform presents the story of a little-known three-year experiment conducted in Japan under American occupation. Beginning in 1948 it sought to test over a period of three years whether schoolchildren could progress faster in a variety of subjects if the written language of instruction was restricted entirely to romaji. Thus, a textbook explanation of a law of physics in romaji obviated the need to recognise or learn kanji. This experimental use of romaji, however, was avoided in Japanese language and literature syllabuses. The results were not conclusive, but the data strongly suggested children in the experimental romaji classes made faster progress than their counterparts under normal instruction.

The inheritance of kanji is one of the subtlest beauties of Japanese writing, but it slows the acquisition of literacy. Today literacy in Japan requires the mastering of 2,000 kanji and recognising another 1,000 or more. In the last century the total was higher, and polite writing style borrowed heavily from Chinese literature for classical forms and phrases. In 1895, during the Shimonoseki peace conference, foreign minister Mutsu Munemitsu's ease in reading Li Hongzhang's memoranda written in literary Chinese led Li to exclaim that their countries shared a common language. This enthusiastic comment, however, recognised only a polar degree of literary aptitude to which few in Japan could aspire. From the late 19th century onwards, reform of the writing system was to become an integral strand of Japan's modernisation.

In an excellent account of literacy, J. Marshall Unger surveys Japanese script reform and its implications for popular literacy from the Meiji period until the end of the Pacific war. Japan was an extremely literate society during this period, but Unger challenges common assumptions that literacy was spread evenly throughout the population. He revises some standard items of early census literature to show that literacy defined for official purposes is far removed from "literacy as a vehicle for full and free participation in society''. At Tokiwa village in 1881, less than 2 per cent of 882 males could read official proclamations and newspaper editorials, 35 per cent were illiterate, while 41 per cent could write their names and addresses but needed help for other writing tasks. Unger ought perhaps also to reveal that Tokiwa is near the top of Hokkaido, culturally as far flung as, say, the Aran islands from 19th-century Dublin. Nevertheless, his reassessment of the census data is highly suggestive. In 1948, only 6 per cent of the Japanese population did not experience some difficulty in reading or writing Japanese.

Small comfort probably in 1940 for most recruits in the imperial army at the mercy of 1,235 kanji for weapons' parts. The romaji experiment is a deeply political story. Although romaji trials attracted strong support among many Japanese linguists and educators, the nationwide experiment was primarily set in motion by an American initiative, from late 1948 onwards. Reactions to its progress were mixed. The most negative cast the experiment's fieldworkers and teachers as cultural iconoclasts, indeed even demonised them as agents of the abolition of all scripts except romaji. Such disaffection presented hazardous repercussions for an occupation government most keen to disengage progressively from Japanese internal affairs. Unger cites hitherto unpublished reports and government documents to relate the struggle between well-intentioned pedagogic inquiry and ruling officialdom's frequently myopic opposition employing all the usual forms of bureaucratic cunning. A report from the experiment's grass roots by Kito Reizo gives a highly evocative picture of how pupils, parents and school principals responded to the experimental classes. It also brings home the stunning boldness of what a small number of people were trying to test in a Japan that was a mere five years on from one of the most disastrous bouts of militarism this century.

Finally, ahead of term, the romaji experiment was cancelled. It is impossible not to choose sides in this story of scientific objectivity versus political expediency, and the book is all the more readable for Unger's obvious sympathy with the experiment's aims and his respect for those who tried to achieve them. Romanisation is and will remain a topic of East Asian significance, and not least cause for the odd surprise. This reviewer could not help noticing during a recent trip to China that his landing card offered space for his name in English and Chinese but requested Japanese arrivals to give their full names only in the "Rome alphabet".

Oliver Moore is curator, department of oriental antiquities, British Museum.